Advanced PD noise suppression and its relevance for computer aided PD defect identification

Digital partial discharge (PD) defect identification has become state of the art. But computer aided procedures based on pattern recognition principles are negatively effected by pulse-shaped disturbances. It is pointed out that a sufficient noise suppression and noise resistivity is a must for automated diagnosis. A novel approach is presented, which improves the defect identification on-site at noisy conditions. The benefits and limits of expert diagnosis compared to machine intelligent systems are discussed. Experimental results are presented, taken into account hard- and software noise suppression components. It is shown that PD defect identification at operating voltages and respectively disturbed data is still a challenge. A method of resolution with an intelligent neural noise filter and noise resistant diagnosis concepts are discussed.